R Notebook

packages = c('rgdal', 'maptools', 'raster','spatstat', 'tmap', 'sf', 'tidyverse', 'SpatialAcc', 'tbart', 'cclust', 'rgeos') 
for (p in packages){ 
  if(!require(p, character.only = T)){ 
    install.packages(p) } 
  library(p,character.only = T) }

Singapore Planning Subzone (MP14_SUBZONE_WEB_PL)

mpsz = st_read(dsn = "data", layer = "MP14_SUBZONE_WEB_PL")

Reading layer `MP14_SUBZONE_WEB_PL' from data source `C:\IS415\IS415-Neighbourhood-WatchDocs\data' using driver `ESRI Shapefile'
Simple feature collection with 323 features and 15 fields
geometry type:  MULTIPOLYGON
dimension:      XY
bbox:           xmin: 2667.538 ymin: 15748.72 xmax: 56396.44 ymax: 50256.33
epsg (SRID):    NA
proj4string:    +proj=tmerc +lat_0=1.366666666666667 +lon_0=103.8333333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +datum=WGS84 +units=m +no_defs

mpsz <- mpsz[order(mpsz$SUBZONE_N),]

mpsz1 <- readOGR(dsn = "data", layer = "MP14_SUBZONE_WEB_PL")

OGR data source with driver: ESRI Shapefile 
Source: "C:\IS415\IS415-Neighbourhood-WatchDocs\data", layer: "MP14_SUBZONE_WEB_PL"
with 323 features
It has 15 fields

mpsz1 

class       : SpatialPolygonsDataFrame 
features    : 323 
extent      : 2667.538, 56396.44, 15748.72, 50256.33  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=tmerc +lat_0=1.366666666666667 +lon_0=103.8333333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0 
variables   : 15
names       : OBJECTID, SUBZONE_NO, SUBZONE_N, SUBZONE_C, CA_IND, PLN_AREA_N, PLN_AREA_C,       REGION_N, REGION_C,          INC_CRC, FMEL_UPD_D,    X_ADDR,   Y_ADDR, SHAPE_Leng,  SHAPE_Area 
min values  :        1,          1, ADMIRALTY,    AMSZ01,      N, ANG MO KIO,         AM, CENTRAL REGION,       CR, 00F5E30B5C9B7AD8, 2014/12/05,  5092.895, 19579.07,   871.5549,    39437.94 
max values  :      323,         17,    YUNNAN,    YSSZ09,      Y,     YISHUN,         YS,    WEST REGION,       WR, FFCCF172717C2EAF, 2014/12/05, 50424.792, 49552.79, 68083.9365, 69748298.79 

subzone <- mpsz[mpsz$SUBZONE_N == "ADMIRALTY",] 
p <- as(subzone, 'Spatial')  
sp_cent <- gCentroid(p, byid = TRUE)
#crs(sp_cent)
# sp_cent <- st_transform(subzone, crs = st_crs(sp_cent))
# 
sp_cent

class       : SpatialPoints 
features    : 1 
extent      : 27091.28, 27091.28, 48333.8, 48333.8  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=tmerc +lat_0=1.366666666666667 +lon_0=103.8333333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0 

#plot(subzone)

Singapore Residents by Subzone, Age Group and Sex (Gender)

population = read_csv("data/respopagsex2000to2017.csv")

Parsed with column specification:
cols(
  PA = col_character(),
  SZ = col_character(),
  AG = col_character(),
  Sex = col_character(),
  Pop = col_double(),
  Time = col_double()
)

Singapore Residents by Subzone, Age Group and Sex, June 2017 (Gender)

popage2017pre <- population %>%
  filter(Time == 2017) %>%
  filter(AG == "65_to_69" | AG == "70_to_74" | AG == "75_to_79" | AG == "80_to_84" | AG == "85_and_over") %>%
  group_by(SZ)
popage2017pre_added <- aggregate(Pop ~ SZ, popage2017pre, sum)

Total Number of Clinics in Singapore

clinics = read_csv("data/clinics_hcidirectory.csv")

Parsed with column specification:
cols(
  clinic_name = col_character(),
  address = col_character(),
  postal_code = col_character(),
  X = col_double(),
  Y = col_double(),
  LAT = col_double(),
  LONG = col_double()
)

clinics_essentials <- c("clinic_name", "address", "LAT", "LONG", "X", "Y") # remove postal code
clinics <- clinics[clinics_essentials]

Total Number of TCM Clinics in Singapore

tcm = read_csv("data/tcm_tcmboard.csv")

Parsed with column specification:
cols(
  tcm_physician_name = col_character(),
  registration_no = col_character(),
  tcm_place_name = col_character(),
  tcm_address = col_character(),
  postal_code = col_character(),
  X = col_double(),
  Y = col_double(),
  LAT = col_double(),
  LONG = col_double()
)

tcm <- tcm[!duplicated(tcm$tcm_place_name), ]
  
tcm_essentials <- c("tcm_place_name", "tcm_address", "LAT", "LONG", "X", "Y")
tcm <- tcm[tcm_essentials]
names(tcm)[names(tcm) == "tcm_place_name"] <- "clinic_name"
names(tcm)[names(tcm) == "tcm_address"] <- "address"

Combine TCM and clinics

clinics_combined <- rbind(clinics, tcm)
clinics_combined <- st_join(st_as_sf(clinics_combined, 
                                 coords = c("X", "Y"), crs = st_crs(mpsz)), mpsz)

Number of HDB blocks per planning subzone

HDB = read_csv("data/hdb_property_information.csv")

Parsed with column specification:
cols(
  blk_no_street = col_character(),
  total_dwelling_units = col_double(),
  `1room_sold` = col_double(),
  `2room_sold` = col_double(),
  `3room_sold` = col_double(),
  `4room_sold` = col_double(),
  `5room_sold` = col_double(),
  exec_sold = col_double(),
  multigen_sold = col_double(),
  studio_apartment_sold = col_double(),
  `1room_rental` = col_double(),
  `2room_rental` = col_double(),
  `3room_rental` = col_double(),
  other_room_rental = col_double(),
  postal_code = col_character(),
  X = col_double(),
  Y = col_double(),
  LAT = col_double(),
  LONG = col_double()
)

HDB_sf <- st_as_sf(HDB, coords = c("X", "Y"), crs = st_crs(mpsz))

Residents by Age Group & Type of Dwelling, Annual

popByDwelling = read_csv("data/residents-by-age-group-and-type-of-dwelling-detailed-categories-annual.csv")

Parsed with column specification:
cols(
  year = col_double(),
  level_1 = col_character(),
  level_2 = col_character(),
  level_3 = col_character(),
  value = col_double()
)

Residents by Age Group & Type of Dwelling, Annual 2017

popByDwelling2017 <- popByDwelling %>%
  filter(year == 2017) %>%
  filter(level_1 == "65-69 Years" 
         | level_1 == "85 Years & Over" 
         | level_1 == "70-74 Years" 
         | level_1 == "75-79 Years"  
         | level_1 == "80-84 Years") %>%
  group_by(level_3)
popByDwelling2017

popByDwelling2017_added <- aggregate(value ~ level_3, popByDwelling2017, sum)

No. of blocks per subzone

mpsz_HDB <- st_join(HDB_sf,mpsz)
mpsz_HDB$`1_2room_total` <- mpsz_HDB$`1room_sold` + mpsz_HDB$`2room_sold` + mpsz_HDB$`1room_rental` + mpsz_HDB$`2room_rental`
mpsz_HDB_1_2_room = aggregate(mpsz_HDB$`1_2room_total`, by=list(SUBZONE=mpsz_HDB$SUBZONE_N), FUN=sum) %>%
  rename('No_of_1_2_room' = 'x') 
mpsz_HDB$`3room_total` <- mpsz_HDB$`3room_sold` + mpsz_HDB$`3room_rental` + mpsz_HDB$`other_room_rental`
mpsz_HDB_3_room_added = aggregate(mpsz_HDB$`3room_total`, by=list(SUBZONE=mpsz_HDB$SUBZONE_N), FUN=sum) %>%
  rename('No_of_3_room' = 'x') 
mpsz_HDB$`4room_total` <- mpsz_HDB$`4room_sold`
mpsz_HDB_4_room_added = aggregate(mpsz_HDB$`4room_total`, by=list(SUBZONE=mpsz_HDB$SUBZONE_N), FUN=sum) %>%
  rename('No_of_4_room' = 'x') 
mpsz_HDB$`5room_exec_total` <- mpsz_HDB$`5room_sold` + mpsz_HDB$`exec_sold`
mpsz_HDB_5_room_exec_added = aggregate(mpsz_HDB$`5room_exec_total`, by=list(SUBZONE=mpsz_HDB$SUBZONE_N), FUN=sum) %>%
  rename('No_of_5_room_exec' = 'x')   
mpsz_HDB_added <- left_join(mpsz_HDB_1_2_room, mpsz_HDB_3_room_added, by ='SUBZONE') %>%
  left_join(., mpsz_HDB_4_room_added, by = 'SUBZONE') %>%
  left_join(., mpsz_HDB_5_room_exec_added, by = 'SUBZONE')
mpsz_HDB_added$`total_units` <- mpsz_HDB_added$`No_of_1_2_room` + mpsz_HDB_added$`No_of_3_room` + mpsz_HDB_added$`No_of_4_room` + mpsz_HDB_added$`No_of_5_room`

total_1_2_room_units <- sum(mpsz_HDB_added$No_of_1_2_room) 
total_3_room_units <- sum(mpsz_HDB_added$No_of_3_room) 
total_4_room_units <- sum(mpsz_HDB_added$No_of_4_room) 
total_5_room_exec_units <- sum(mpsz_HDB_added$No_of_5_room_exec) 

No. of elderly per block in every subzone

total_1_2_room_units <- sum(mpsz_HDB_added$No_of_1_2_room) 
total_3_room_units <- sum(mpsz_HDB_added$No_of_3_room) 
total_4_room_units <- sum(mpsz_HDB_added$No_of_4_room) 
total_5_room_units <- sum(mpsz_HDB_added$No_of_5_room) 
mpsz_HDB$No_of_Elderly_in_block_1_2 <- ifelse(mpsz_HDB$`1_2room_total` == 0, 0, mpsz_HDB$`1_2room_total`/total_1_2_room_units) * popByDwelling2017_added$value[popByDwelling2017_added$level_3=="HDB 1- And 2-Room Flats"]
mpsz_HDB$No_of_Elderly_in_block_3 <- ifelse(mpsz_HDB$`3room_total` == 0, 0, mpsz_HDB$`3room_total`/total_3_room_units) * popByDwelling2017_added$value[popByDwelling2017_added$level_3=="HDB 3-Room Flats"] 
mpsz_HDB$No_of_Elderly_in_block_4 <- ifelse(mpsz_HDB$`4room_total` == 0, 0, mpsz_HDB$`4room_total`/total_4_room_units) * popByDwelling2017_added$value[popByDwelling2017_added$level_3=="HDB 4-Room Flats"]
mpsz_HDB$No_of_Elderly_in_block_5 <- ifelse(mpsz_HDB$`5room_exec_total` == 0, 0, mpsz_HDB$`5room_exec_total`/total_5_room_exec_units) * popByDwelling2017_added$value[popByDwelling2017_added$level_3=="HDB 5-Room And Executive Flats"]
                                    
mpsz_HDB$No_of_Elderly_in_block <- mpsz_HDB$No_of_Elderly_in_block_1_2 + mpsz_HDB$No_of_Elderly_in_block_3 + mpsz_HDB$No_of_Elderly_in_block_4 + mpsz_HDB$No_of_Elderly_in_block_5

mpsz_HDB$No_of_Elderly_in_block <- round(mpsz_HDB$No_of_Elderly_in_block)

#assign capacity of 2 per clinic
clinics_combined['capacity'] <- 2

mpsz_HDB_coords <- mpsz_HDB %>% st_coordinates() 
clinics_combined_coords <- clinics_combined %>% st_coordinates()

dm <- distance(mpsz_HDB_coords, clinics_combined_coords)
acc_hansen <- data.frame(ac(mpsz_HDB$No_of_Elderly_in_block, clinics_combined$capacity, dm, power=0.01, family="Hansen"))
colnames(acc_hansen) <- "accHansen"
acc_hansen <- tbl_df(acc_hansen)
HDB_acc <- bind_cols(mpsz_HDB, acc_hansen)

tmap_mode("view")

tmap mode set to interactive viewing

tm_shape(mpsz_HDB)+
  tm_symbols(size = 0.1)+
tm_shape(HDB_acc)+
tm_bubbles(col = "accHansen",
           n=5,
           style = "quantile",
           size = 0.1,
           border.col = "black",
           border.lwd = 1)

mpsz_HDB_filtered <- mpsz_HDB[mpsz_HDB$SUBZONE_N=="ADMIRALTY", ]
clinics_combined_filtered <- clinics_combined[clinics_combined$SUBZONE_N=="ADMIRALTY", ]
clinics_combined_filtered <- na.omit(clinics_combined_filtered)

mpsz_HDB_filtered_sp <- as_Spatial(mpsz_HDB_filtered)
clinics_combined_filtered_sp <- as_Spatial(clinics_combined_filtered)

model1 <- allocations(mpsz_HDB_filtered_sp, clinics_combined_filtered_sp, p=1)
head(model1)

sum(model1$allocdist)

[1] 13308.25

optimal_loc <- unique(model1$allocation)
optimal_loc <- clinics_combined_filtered_sp[optimal_loc, ]
head(optimal_loc)

optimal_loc <- mpsz_HDB_filtered_sp[optimal_loc, ]
head(optimal_loc)

star.model1 <- star.diagram(mpsz_HDB_filtered_sp, clinics_combined_filtered_sp, alloc = model1$allocation)